Can someone in the know tell me whether this should be ignored like most nutritional research?
Some things I noticed: This is a meta-analysis, not a direct study by the authors. This is looking for an effect in other people's data.
There are 18 studies with a total of 1105 individuals. This is very very low for a meta-analysis.
They identified 81 studies that they 'wanted' to use, but they had to exclude all but 18.
They have a nice chart: https://academic.oup.com/view-large/figure/381075699/nuac086...
Look at the rightmost 2 columns. The only 'good' outcome with a large effect has low certainty. (Bottom Row, NAFLD)
The only outcome with at least moderate effect and certainty has a bad outcome (Inflammation markers)
In summary, the press release is truly garbage.
The study is not bad, it shows the state of knowledge and research in this area, which shows that honey is food, not magic.
1) Mechanistic data we don't fully understand the implications of.
This basically means we observe mechanisms that we believe ought to have some effect on people, and yet when longitudinal studies are done to measure actual long term effects, they are not statistically significant
2) Confounding factors.
In longitudinal studies, it can be very difficult to fully control for confounding factors, as diets are very complex. Take a hypothetical red meat study, where participants are simply instructed to eat "more red meat". Is a McDonalds burger different than a lean, grassfed steak? That's an easy one to control for, not the best example. But basically exemplifying how there can be a lot of variability within the intervention itself when it comes to nutrition